Computing device shared resource lock allocation

A computing device comprising:

• a plurality of central processing units (CPUs) comprising a first CPU and a second CPU;
• and
• a plurality of node controllers (NCs) in a one-to-one correspondence with the CPUs, wherein the NCs comprise a first NC associated with the first CPU, a second NC, and a third NC associated with the second CPU,
• wherein the first CPU is configured to:
• send, to the first NC, a third lock request requesting a first target lock;
• query a third lock flag bit of the third lock request in a first local queue of the first NC;
• access a shared resource corresponding to the first target lock when the third lock flag bit indicates a locked state;
• and
• send, to the first NC, a request for releasing the first target lock after accessing the shared resource,
• wherein the first NC is configured to:
• receive the third lock request from the first CPU;
• add the third lock request to the first local queue;
• add a first identifier of the first NC to the third lock request to obtain a first lock request;
• and
• send the first lock request to the second NC managing the first target lock, wherein the first local queue is configured to store a lock request from the first CPU, and wherein a remote queue of the first NC and associated with the first target lock is configured to store at least one lock request for requesting the first target lock,
• wherein the second NC is configured to:
• receive the first lock request;
• add the first lock request to the remote queue;
• remove the at least one lock request from the remote queue to generate a first migration queue, wherein a lock request at a first queue head in the first migration queue is the first lock request, and wherein the first lock request comprises the first identifier;
• change a first lock flag bit of the first lock request to the locked state;
• and
• send the first migration queue to the first NC,
• wherein the first NC is further configured to:
• receive the first migration queue;
• allocate the first target lock to the first CPU;
• delete the first lock request at the first queue head when receiving the request from the first CPU;
• change a second lock flag bit of a second lock request to the locked state when a to-be-processed lock request exists in the first migration queue from which the first lock request is deleted, wherein the second lock request is at the first queue head in the first migration queue from which the first lock request is deleted, wherein the locked state indicates that the first target lock is occupied, wherein the second lock request comprises a third identifier of the third NC sending the second lock request;
• and
• send the first migration queue to the third NC, and
• wherein the third NC is configured to:
• receive the first migration queue from the first NC;
• and
• change, according to the second lock flag bit, a lock flag bit of a lock request corresponding to the second lock request in a third local queue of the third NC and storing a lock request from the second CPU.

A first node controller (NC) comprising:

• a storage circuit configured to store instructions;
• and
• a processor coupled to the storage circuit and configured to execute the instructions to:
• receive, from a second NC managing a first target lock, a first migration queue comprising at least one lock request for requesting the first target lock, wherein a first lock request is at a first queue head in the first migration queue and comprises a first identifier of the first NC;
• allocate the first target lock to a first central processing unit (CPU) associated with the first NC;
• delete the first lock request when receiving a request for releasing the first target lock from the first CPU;
• change a second lock flag bit of a second lock request to a locked state when a to-be-processed lock request exists in the first migration queue from which the first lock request is deleted, wherein the locked state indicates that the first target lock is occupied, wherein the second lock request is at the first queue head in the first migration queue from which the first lock request is deleted, and wherein the second lock request comprises a third identifier of a third NC sending the second lock request;
• send a response packet to the second NC when the first migration queue from which the first lock request is deleted is empty, wherein the response packet indicates that all lock requests in the first migration queue are processed, wherein the lock requests comprise the at least one lock request;
• and
• send the first migration queue to the third NC to instruct the third NC to change, based on the second lock flag bit, a lock flag bit of a lock request corresponding to the second lock request in a third local queue of the third NC and storing a lock request from a CPU associated with the third NC.